How the Awesomeness Score works.
No black box - the Awesomeness Score is a weighted audit across technology choice, version health and engineering quality. Published, reproducible, and honest about its limits.
Three audited dimensions
One score, three inputs. Each is measured on its own terms before they are weighted together.
How good are the choices?
Every detected technology is ranked inside its own category by real signals - npm download volume and GitHub reputation. A popular, well-maintained framework scores high; a niche or abandoned one drags the category down. We measure the quality of the choices, not how many boxes you tick.
How current are the versions?
Each detected version is placed on its own release timeline. Running the latest release scores full marks; falling several versions behind costs points, because old versions are where known vulnerabilities and missing fixes live.
How well is it built?
A Lighthouse audit, weighted the way an engineer would: accessibility and best practices count ×10, PWA ×5, performance and SEO ×1. Raw speed can't paper over structural debt - a fast page that fails accessibility still loses points.
How the weighting works
The three dimensions don't count equally - and neither do good and bad sub-scores. AwesomeTechStack weights weak results far more heavily than strong ones, because that is how real risk behaves.
One outdated, vulnerable dependency hurts your score more than ten great choices help it. That is deliberate: the score should reward fixing the worst thing first.
Reading the score
The fourth dimension: AI-Readiness
llms.txt, agent access rules, structured data, content without JavaScript - graded A-F on new audits as the checks ship. It is displayed alongside the Awesomeness Score, not weighted into it yet. When the AI web settles, it joins the weighting - announced right here.
How we grade it ->